Fault tolerant storage systems may split a data fragment into data chunks. The system may also generate a number of additional chunks on the basis of the data chunks. The data chunks and additional chunks may be stored on a number of servers. The number of servers may equal the number of data chunks plus the number of additional chunks. The data fragment may be considered saved when the servers confirm that each data chunk and each additional chunk was successfully saved. This process may be known as archiving and may take a very long time. For example, if one of the servers is slow or unavailable at a particular time, it may take a long time for that server to confirm that a chunk was successfully saved, and the archiving process may ultimately fail.
The present application addresses this need and others.